Passage Reasoning: Evidence, Scope and Uncertainty
Learn to separate supported claims, contradictions and missing information; review both verbal formats with fresh examples and a practical evidence checklist.
Read this full public learning guide before or after practice. It is identical for everyone. Your private on-page record separately shows correct, wrong and unanswered counts out of six; no percentile, employer cutoff or ability level is inferred.
Read the response rule first
- This part deliberately contains two response formats. For the first three items, decide whether a statement is True, False or Cannot say using only the passage. For the next three, choose the one supported claim among four options. Before reading quickly, identify the requested action. A familiar phrase such as “reasoning test” does not tell you whether you must classify a statement, select an inference or obey an instruction.
- The bank is English in both interfaces; the Chinese interface translates guidance and explanations rather than offering a second measured verbal form. An unfamiliar word, language fluency, screen reading or fatigue can affect this short practice. Keep those conditions in your notes. Six answers do not isolate a general language ability, establish workplace communication competence or rank you against a reference population.
True, False and Cannot say
- True means the passage supplies enough evidence for the particular statement. False means it supplies evidence that conflicts with that statement. Cannot say means the supplied information leaves the statement unresolved. It is not the same as personally lacking confidence. A report can establish that a rule forbids email reservations, making “email is enough” false, even if you have never encountered that booking system in real life.
- Try a fresh example: a desk records 18 enquiries in June and 24 in July but gives no opening hours. “July had more recorded enquiries” is supported. “June had more recorded enquiries” conflicts with the totals. “July had a higher hourly enquiry rate” is unresolved. Adding plausible working hours from your experience would change the task instead of resolving the missing denominator from the text.
Match scope and qualifiers
- Underline the population, period and strength of each claim. “Some volunteers attended” cannot by itself establish that most employees prefer the event. “Only after approval” states a necessary condition, not a promise that every approved request will happen. Words such as all, always, only and at least change the claim. Compare their exact scope with the passage rather than selecting the option that sounds most generally reasonable.
- For example, a trial may describe results for participants who signed up, without describing everyone eligible. Keep that selection boundary visible. Do not turn an observation in one room into a universal recommendation, or confuse a request to investigate a delay with proof that a delay has occurred. A useful check is to copy the smallest phrase that justifies each important word in your candidate answer.
Supported claims and causal restraint
- In the four-option format, reject options that add facts, change a condition or overstate causation. If two rooms differ in labels, training and layout, an observed error difference does not isolate label colour. This does not prove that colour is irrelevant; it says the study as described does not separate its effect. Prefer the bounded conclusion actually justified, rather than the strongest interesting explanation you can imagine.
- A second fresh example: two teams switch both software and shift schedules, after which response times fall. “The combined change was followed by faster responses” may describe the record. “The software alone caused the improvement” requires evidence not provided. When reviewing your choice, name the extra assumption rather than merely saying that an option felt too strong. That makes the error usable in another task.
Use the short section clock deliberately
- Timed practice gives this six-item part three minutes, a site-selected cap rather than an official pace. A rough thirty seconds per item is only a planning average; it is not an instruction to spend equal time on unequal passages. Read the response rule, locate the decisive evidence, choose or explicitly leave a blank, and move on. Backtracking remains possible within the active part until it closes.
- The next part waits for your Start button, so use the between-part pause to reset your attention. An active clock continues in the background. Untimed study is appropriate for checking how your inference follows the words before adding time pressure. A missed deadline is a local practice event, not evidence about an employer cutoff. For an actual assessment, read its own timing, navigation and accessibility instructions.
Review the first unsupported step
- After all three parts close, open every explanation, including questions you got right. For a wrong selection, mark the first move that left the passage: adding a population, assuming equal hours, reversing a condition or claiming one cause. For a correct selection, check that your reason matches the evidence rather than an accidental guess. Keep wrong selections and unanswered items separate when choosing what to practise next.
- Write a new two-sentence passage and one claim for each of the three classifications. Then change just one condition and see whether the classification changes. An example that stays ambiguous after adding evidence may reveal a vague statement rather than a reasoning failure. This small construction exercise tests your rule understanding more directly than memorising that a particular fixed-bank question uses option C.
Interpret repeats and public guidance
- Your private record shows raw counts out of six, not a calibrated verbal score. Three-choice and four-choice questions have different response sets; neither has been equated for difficulty. A larger count on a repeat may reflect recognition, changed reading conditions or better reasoning. This practice cannot separate those contributions, and a lower count than the other parts does not establish your weakest underlying ability.
- This entire guide is public and identical for everyone. It contains no saved answers, private attempt link or personalised hiring recommendation. Sharing the public URL shares only learning material. If you want more reading practice, use fresh passages under a clear response rule and review the evidence for each answer. Provider preparation material and your exact invitation remain the authority for the assessment you will actually take.